At present, a corresponding text on a webpage is recommended to a user by generating a relevant text at the bottom of the webpage manually or automatically.
In an existing method for recommending a text, the similarity between a new text and an original text is calculated when the corresponding text is generated. The time required for calculating will grow with accumulation of the text. This will cause information overload and the text can not be recommended to the user timely.
Additionally, an appropriate information dimension is not adopted in prior art, so that the function of subscribing critical information can not be offered to the user and the effect of recommending text information is reduced.